Abstract

Research on linguistic diversity has revived a number of debates concerning universal and language-specific determinants in language acquisition. The present paper addresses some of these questions in relation to the expression of caused motion. In this domain languages fall into two typologically different families (Talmy, 2000), in which 1) Manner/Cause are typically encoded in verbal roots and Path in satellites (to walk, hop, skip… roll the ball… across, up, down…), or 2) Path is expressed in main verbs and information generally encoded in less compact constructions (traverser en courant ‘to run across’, traverser en faisant-rouler ‘to go across by making-roll’).

We will examine the implications of this typological contrast for second language acquisition. Several groups of adult speakers (English native speakers, French native speakers – and English learners of French at three proficiency levels) were asked to describe animated cartoons in which an agent acted upon an object in a certain Manner causing its displacement according to a certain Manner and Path (e.g., push a ball so that it rolls down a hill). At lower proficiency levels some of the learners’ responses relied on Path verbs, but did not explicitly express all of the information (e.g., intransitive entrer ‘enter’). Other responses relied on Manner verbs with satellites that unsuccessfully attempted to indicate location changes (e.g., marcher/pousser dans ‘to walk/push in’, marcher à travers ‘to walk across’). With increasing proficiency speakers used more complex constructions, but whereas French natives expressed Path in main verbs and Manner in gerunds (e.g., monter en poussant ‘to ascend pushing’), learners frequently did the reverse (e.g. pousser en montant ‘to push ascending’). Regardless of proficiency, many responses clearly had a non-native flavour, despite the fact that they were not ungrammatical. Such results constitute a real challenge for language teachers. More generally, the discussion highlights the implications of typological constraints for models of second language acquisition and teaching.